dba基礎課程：Oracle備份恢復-手動不完全恢復
平臺：redhat linux as5,oracle10g
不完全恢復是指把數據庫恢復到備份點和失敗點之間某個時刻的狀態,而且不完全恢復只適用於archivelog模式,不完全恢復可以基於時間點,基於SCN,基於控制文件或基於取消.下面用實例演示一下四種不完全恢復的用法.
一.基於時間的不完全恢復
一般當用戶誤刪除表,誤載斷表,提交了錯誤數據後,DBA可以基於時間把數據庫恢復到提交數據前的某一個狀態
現在數據庫中有一個表chenxy
SQL> select * from chenxy;
I
----------------------
10
20
先做個全備份(數據文件)
SQL> shutdown immediate;
SQL> ! cp /u01/oradata/denver/*.dbf /u01/backup
現看當前時間並載斷表
SQL> select to_char(sysdate,'yyyy-mm-dd hh24:mi:ss') from dual;
TO_CHAR(SYSDATE,'YY
-------------------
2009-05-05 23:23:11
SQL> truncate table t;
Table truncated.
現在我們把表chenxy恢復到2009-05-05 23:23:11
1.關閉數據庫
SQL> shutdown immediate;
Database closed.
Database dismounted.
ORACLE instance shut down.
2.裝載數據庫
SQL> startup mount;
ORACLE instance started.
Total System Global Area 420549952 bytes
Fixed Size 451904 bytes
Variable Size 385875968 bytes
Database Buffers 33554432 bytes
Redo Buffers 667648 bytes
Database mounted.
3.複製所有數據文件備份
SQL> ! cp /u01/backup/*.dbf /u01/oradata/denver/
注意備份的時間點要在2009-05-05 23:23:11之前,查看備份時間
SQL> select file#,to_char(time,'yyyy-mm-dd hh24:mi:ss') from v$recover_file;
FILE# TO_CHAR(TIME,'YYYY-
------------- -------------------
1 2009-05-05 23:20:18
2 2009-05-05 23:20:18
3 2009-05-05 23:20:18
4 2009-05-05 23:20:18
5 2009-05-05 23:20:18
6 2009-05-05 23:20:18
4.恢復到特定時間點
SQL> recover database until time '2009-05-05 23:23:24';
Media recovery complete.
5.以resetlogs方式打開數據庫,並查看錶chenxy
SQL> alter database open resetlogs;
Database altered.
SQL> select * from chenxy;
I
--------------------
10
20
當以resetlogs方式打開數據庫時,會重新建立重做日誌,清空原有重做日誌的內容(同時歸檔日誌也全部刪除了),而且過去的備份也不能直接使用了,現在重新備份所有數據文件和控制文件,並歸檔當前日誌組
SQL> shutdown immediate;
SQL> ! cp /u01/oradata/denver/*.dbf /u01/backup
SQL> ! cp /u01/oradata/denver/*.ctl /u01/backup
SQL>startup
SQL> alter system switch logfile;
6.簡單排錯
因爲選項resetlogs要清空重做日誌,當出現如下錯誤ORA-00338時可用resetlogs打開數據庫
SQL> startup
ORACLE instance started.
Total System Global Area 420549952 bytes
Fixed Size 451904 bytes
Variable Size 385875968 bytes
Database Buffers 33554432 bytes
Redo Buffers 667648 bytes
Database mounted.
ORA-00338: log 2 of thread 1 is more recent than controlfile
ORA-00312: online log 2 thread 1: '/u01/oradata/denver/redo02.log'
使用resetlogs打開後一定要做備份.
SQL> alter database open resetlogs;
Database altered.
SQL> select status from v$instance;
STATUS
------------
OPEN
二.基於SCN恢復
1.查看當前SCN號
SQL> select current_scn from v$database;
#或
SQL> select dbms_flashback.get_system_change_number scn from dual;
SCN
--------------------
1673513
1.關閉數據庫
SQL> shutdown immediate;
Database closed.
Database dismounted.
ORACLE instance shut down.
2裝載數據庫
SQL> startup mount;
ORACLE instance started.
Total System Global Area 420549952 bytes
Fixed Size 451904 bytes
Variable Size 385875968 bytes
Database Buffers 33554432 bytes
Redo Buffers 667648 bytes
Database mounted.
3.複製所有數據文件備份
SQL> ! cp /u01/backup/*.dbf /u01/oradata/denver/
注意備份的SCN號要小於1673513,查看備份的SCN
SQL> select file#,change# from v$recover_file;
FILE# CHANGE#
---------- ----------
1 1673087
2 1673087
3 1673087
4 1673087
5 1673087
6 1673087
7 1673087
4.恢復到特定SCN
SQL> recover database until change 1673513;
Media recovery complete.
5.以resetlogs方式打開數據庫,並查看錶chenxy
SQL> alter database open resetlogs;
Database altered.
SQL> select * from chenxy;
I
--------------------
10
20
6.重新備份所有數據文件和控制文件,並歸檔當前日誌組
SQL> shutdown immediate;
SQL> ! cp /u01/oradata/denver/*.dbf /u01/backup
SQL> ! cp /u01/oradata/denver/*.ctl /u01/backup
SQL>startup
SQL> alter system switch logfile;
三.基於取消不完全恢復
其於取消恢復主要適用於因歸檔日誌丟失,只能恢復到某一個時間的情況,恢復方法和前面兩種一樣,只是恢復時點不同.
恢復
SQL>shutdown immediate;
SQL>startup mount;
SQL>! cp /u01/backup/*.dbf /u01/oradata/denver/
SQL> recover database until cancel;
Media recovery complete.
SQL>alter database open resetlogs;
四.基於備份控制文件恢復
原理:當用戶誤刪除一個表空間時,當前的控制文件就把此表空間的信息都刪了,但是在備份的控制文件裏還是它的信息,所以可以用備份的控制文件恢復表空間
刪除一個表空間
SQL> drop tablespace chenxy including contents;
Tablespace dropped.
SQL> select * from chenxy;
select * from chenxy
ERROR at line 1:
ORA-00942: table or view does not exist
到alter日誌找到刪除時間
SQL> show parameter background_dump_dest
NAME TYPE VALUE
------------------------------------ ----------- ------------------------------
background_dump_dest string /u01/admin/denver/bdump
SQL>! less /u01/admin/denver/bdump/alert_denver.log
Wed May 6 00:10:19 2009
drop tablespace chenxy including contents
Completed: drop tablespace chenxy including contents
恢復
SQL>shutdown immediate;
SQL>startup mount;
#此時把控制文件也複製過來
SQL>! cp /u01/backup/*.dbf /u01/oradata/denver/
SQL>! cp /u01/backup/*.bak /u01/oradata/denver/
SQL>recover database until time '2009-05-05 23:50:24' using backup controlfile
SQL>alter database open resetlogs
最後最好再做一個全備份.
來源：Oracle培訓
